Using a Python dict for a SQL INSERT statement
Question:
I am trying to use a dict
to do a SQL INSERT
. The logic would basically be:
INSERT INTO table (dict.keys()) VALUES dict.values()
However, I am having a tough time figuring out the correct syntax / flow to do this. This is what I currently have:
# data = {...}
sorted_column_headers_list = []
sorted_column_values_list = []
for k, v in data.items():
sorted_column_headers_list.append(k)
sorted_column_values_list.append(v)
sorted_column_headers_string = ', '.join(sorted_column_headers_list)
sorted_column_values_string = ', '.join(sorted_column_values_list)
cursor.execute("""INSERT INTO title (%s)
VALUES (%s)""",
(sorted_column_headers_string, sorted_column_values_string))
From this I get a SQL exception (I think related to the fact that commas are also included in some of the values that I have). What would be the correct way to do the above?
Answers:
You want to add parameter placeholders to the query. This might get you what you need:
qmarks = ', '.join('?' * len(myDict))
qry = "Insert Into Table (%s) Values (%s)" % (qmarks, qmarks)
cursor.execute(qry, myDict.keys() + myDict.values())
I think the comment on using this with MySQL is not quite complete. MySQLdb doesn’t do parameter substitution in the columns, just the values (IIUC) – so maybe more like
placeholders = ', '.join(['%s'] * len(myDict))
columns = ', '.join(myDict.keys())
sql = "INSERT INTO %s ( %s ) VALUES ( %s )" % (table, columns, placeholders)
# valid in Python 2
cursor.execute(sql, myDict.values())
# valid in Python 3
cursor.execute(sql, list(myDict.values()))
You’re not getting escaping on the columns though, so you might want to check them first….
See http://mail.python.org/pipermail/tutor/2010-December/080701.html for a more complete solution
table='mytable'
columns_string= '('+','.join(myDict.keys())+')'
values_string = '('+','.join(map(str,myDict.values()))+')'
sql = """INSERT INTO %s %s
VALUES %s"""%(table, columns_string,values_string)
What about:
keys = str(dict.keys())
keys.replace('[', '(')
keys.replace(']', ')')
keys.replace("'",'')
vals = str(dict.values())
vals.replace('[', '(')
vals.replace(']', ')')
cur.execute('INSERT INTO table %s VALUES %s' % (keys, vals))
For python 3:
keys = str(dict.keys())[9:].replace('[', '').replace(']', '')
vals = str(dict.values())[11:].replace('[', '').replace(']', '')
…
Always good answers here, but in Python 3, you should write the following:
placeholder = ", ".join(["%s"] * len(dict))
stmt = "insert into `{table}` ({columns}) values ({values});".format(table=table_name, columns=",".join(dict.keys()), values=placeholder)
cur.execute(stmt, list(dict.values()))
Don’t forget to convert dict.values()
to a list because in Python 3, dict.values()
returns a view, not a list.
Also, do NOT pour the dict.values()
in stmt
because it tears a quote out of a string by join
ing it, which caused MySQL error in inserting it. So you should always put it in cur.execute()
dynamically.
I tried @furicle’s solution but it still inputs everything as a string – if your dict is a mixed one then this may not work as you would want it to. I had a similar issue and this is what I came up with – this is only a query builder and you could use it (with changes) to work with any database of your choice. Have a look!
def ins_query_maker(tablename, rowdict):
keys = tuple(rowdict)
dictsize = len(rowdict)
sql = ''
for i in range(dictsize) :
if(type(rowdict[keys[i]]).__name__ == 'str'):
sql += ''' + str(rowdict[keys[i]]) + '''
else:
sql += str(rowdict[keys[i]])
if(i< dictsize-1):
sql += ', '
query = "insert into " + str(tablename) + " " + str(keys) + " values (" + sql + ")"
print(query) # for demo purposes we do this
return(query) #in real code we do this
This is crude and still needs sanity checks, etc, but it works as intended.
for a dict:
tab = {'idnumber': 1, 'fname': 'some', 'lname': 'dude', 'dob': '15/08/1947', 'mobile': 5550000914, 'age' : 70.4}
running the query I get the following output
I used this thread for my usage and tried to keep it much simpler
ins_qry = "INSERT INTO {tablename} ({columns}) VALUES {values};" .format(
tablename=my_tablename,
columns=', '.join(myDict.keys()),
values=tuple(myDict.values())
)
cursor.execute(ins_qry)
Make sure to commit the data inserted, either using db_connection.commit()
and use cursor.lastrowid
, if you need the primary key of the inserted row
columns = ', '.join(str(x).replace('/', '_') for x in row_dict.keys())
values = ', '.join("'" + str(x).replace('/', '_') + "'" for x in row_dict.values())
sql = "INSERT INTO %s ( %s ) VALUES ( %s );" % ("tablename", columns, values)
applicable for python3
This code worked for me (Python 3):
fields = (str(list(dictionary.keys()))[1:-1])
values = (str(list(dictionary.values()))[1:-1])
sql = 'INSERT INTO Table (' + fields + ') VALUES (' + values + ')'
cursor.execute(sql)
It does rely on the dictionary outputting its keys and values in the same order. I’m unclear if this is always true 🙂
I’m a little late to the party but there is another way that I tend to prefer since my data is usually in the form of a dict already. If you list the bind variables in the form of %(columnName)s you can use a dictionary to bind them at execute. This partially solves the problem of column ordering since the variables are bound in by name. I say partially because you still have to make sure that the columns & values portion of the insert are mapped correctly; but the dictionary itself can be in any order (since dicts are sort of unordered anyway)
There is probably a more pythonic way to achieve all this, but pulling the column names into a list and working off it ensures we have a static ordering to build the columns & values clauses.
data_dict = {'col1': 'value 1', 'col2': 'value 2', 'col3': 'value 3'}
columns = data_dict.keys()
cols_comma_separated = ', '.join(columns)
binds_comma_separated = ', '.join(['%(' + item + ')s' for item in columns])
sql = f'INSERT INTO yourtable ({cols_comma_separated}) VALUES ({binds_comma_separated})'
cur.execute(sql, data_dict)
Now whether or not it is a good idea to dynamically build your columns & values clause like this is a topic for a SQL injection thread.
This works for me
cursor.execute("INSERT INTO table (col) VALUES ( %(col_value) )",
{'col_value': 123})
if you have list in which there are number of dictionaries
for example: lst=[d1,d2,d3,d4]
then below one will worked for me:
for i in lst:
placeholders = ', '.join(['%s'] * len(i))
columns = ', '.join(i.keys())
sql = "INSERT INTO %s ( %s ) VALUES ( %s )" % (table, columns, placeholders)
cursor.execute(sql,list(i.values()))
conn.commit()
Note:Dont ever forget to commit otherwise you wont be able to see columns and values inserted in table
When constructing queries dynamically it’s important to ensure that both identifiers and values are correctly quoted. Otherwise you risk
- SQL injection if untrusted data is processed
- Errors if the column names require quoting (for example embedded spaces)
- Data corruption or errors if values are incorrectly quoted (for example
2021-07-11
unquoted may be evaluated as 2003
)
Quoting values is best delegated to the DB-API connector. However connector packages don’t always provide a way to quote identifiers, so you may need to do this manually. MySQL uses backticks (`) to quote identifiers.
This code quotes identifiers and values. It works for MySQLdb, mysql.connector and pymysql and works for Python 3.5+.
data = {'col1': val1, 'col2': val2, ...}
# Compose a string of quoted column names
cols = ','.join([f'`{k}`' for k in data.keys()])
# Compose a string of placeholders for values
vals = ','.join(['%s'] * len(data))
# Create the SQL statement
stmt = f'INSERT INTO `tbl` ({cols}) VALUES ({vals})'
# Execute the statement, delegating the quoting of values to the connector
cur.execute(stmt, tuple(data.values()))
Let’s say our data is:
data = {
"name" : "fani",
"surname": "dogru",
"number" : 271990
}
This is my shorter version:
tablo = "table_name"
cols = ','.join([f" {k}" for k in data.keys()])
vals = ','.join([f"'{k}'" for k in data.values()])
stmt = f'INSERT INTO {tablo} ({cols}) VALUES ({vals})'
I am trying to use a dict
to do a SQL INSERT
. The logic would basically be:
INSERT INTO table (dict.keys()) VALUES dict.values()
However, I am having a tough time figuring out the correct syntax / flow to do this. This is what I currently have:
# data = {...}
sorted_column_headers_list = []
sorted_column_values_list = []
for k, v in data.items():
sorted_column_headers_list.append(k)
sorted_column_values_list.append(v)
sorted_column_headers_string = ', '.join(sorted_column_headers_list)
sorted_column_values_string = ', '.join(sorted_column_values_list)
cursor.execute("""INSERT INTO title (%s)
VALUES (%s)""",
(sorted_column_headers_string, sorted_column_values_string))
From this I get a SQL exception (I think related to the fact that commas are also included in some of the values that I have). What would be the correct way to do the above?
You want to add parameter placeholders to the query. This might get you what you need:
qmarks = ', '.join('?' * len(myDict))
qry = "Insert Into Table (%s) Values (%s)" % (qmarks, qmarks)
cursor.execute(qry, myDict.keys() + myDict.values())
I think the comment on using this with MySQL is not quite complete. MySQLdb doesn’t do parameter substitution in the columns, just the values (IIUC) – so maybe more like
placeholders = ', '.join(['%s'] * len(myDict))
columns = ', '.join(myDict.keys())
sql = "INSERT INTO %s ( %s ) VALUES ( %s )" % (table, columns, placeholders)
# valid in Python 2
cursor.execute(sql, myDict.values())
# valid in Python 3
cursor.execute(sql, list(myDict.values()))
You’re not getting escaping on the columns though, so you might want to check them first….
See http://mail.python.org/pipermail/tutor/2010-December/080701.html for a more complete solution
table='mytable'
columns_string= '('+','.join(myDict.keys())+')'
values_string = '('+','.join(map(str,myDict.values()))+')'
sql = """INSERT INTO %s %s
VALUES %s"""%(table, columns_string,values_string)
What about:
keys = str(dict.keys())
keys.replace('[', '(')
keys.replace(']', ')')
keys.replace("'",'')
vals = str(dict.values())
vals.replace('[', '(')
vals.replace(']', ')')
cur.execute('INSERT INTO table %s VALUES %s' % (keys, vals))
For python 3:
keys = str(dict.keys())[9:].replace('[', '').replace(']', '')
vals = str(dict.values())[11:].replace('[', '').replace(']', '')
…
Always good answers here, but in Python 3, you should write the following:
placeholder = ", ".join(["%s"] * len(dict))
stmt = "insert into `{table}` ({columns}) values ({values});".format(table=table_name, columns=",".join(dict.keys()), values=placeholder)
cur.execute(stmt, list(dict.values()))
Don’t forget to convert dict.values()
to a list because in Python 3, dict.values()
returns a view, not a list.
Also, do NOT pour the dict.values()
in stmt
because it tears a quote out of a string by join
ing it, which caused MySQL error in inserting it. So you should always put it in cur.execute()
dynamically.
I tried @furicle’s solution but it still inputs everything as a string – if your dict is a mixed one then this may not work as you would want it to. I had a similar issue and this is what I came up with – this is only a query builder and you could use it (with changes) to work with any database of your choice. Have a look!
def ins_query_maker(tablename, rowdict):
keys = tuple(rowdict)
dictsize = len(rowdict)
sql = ''
for i in range(dictsize) :
if(type(rowdict[keys[i]]).__name__ == 'str'):
sql += ''' + str(rowdict[keys[i]]) + '''
else:
sql += str(rowdict[keys[i]])
if(i< dictsize-1):
sql += ', '
query = "insert into " + str(tablename) + " " + str(keys) + " values (" + sql + ")"
print(query) # for demo purposes we do this
return(query) #in real code we do this
This is crude and still needs sanity checks, etc, but it works as intended.
for a dict:
tab = {'idnumber': 1, 'fname': 'some', 'lname': 'dude', 'dob': '15/08/1947', 'mobile': 5550000914, 'age' : 70.4}
running the query I get the following output
I used this thread for my usage and tried to keep it much simpler
ins_qry = "INSERT INTO {tablename} ({columns}) VALUES {values};" .format(
tablename=my_tablename,
columns=', '.join(myDict.keys()),
values=tuple(myDict.values())
)
cursor.execute(ins_qry)
Make sure to commit the data inserted, either using db_connection.commit()
and use cursor.lastrowid
, if you need the primary key of the inserted row
columns = ', '.join(str(x).replace('/', '_') for x in row_dict.keys())
values = ', '.join("'" + str(x).replace('/', '_') + "'" for x in row_dict.values())
sql = "INSERT INTO %s ( %s ) VALUES ( %s );" % ("tablename", columns, values)
applicable for python3
This code worked for me (Python 3):
fields = (str(list(dictionary.keys()))[1:-1])
values = (str(list(dictionary.values()))[1:-1])
sql = 'INSERT INTO Table (' + fields + ') VALUES (' + values + ')'
cursor.execute(sql)
It does rely on the dictionary outputting its keys and values in the same order. I’m unclear if this is always true 🙂
I’m a little late to the party but there is another way that I tend to prefer since my data is usually in the form of a dict already. If you list the bind variables in the form of %(columnName)s you can use a dictionary to bind them at execute. This partially solves the problem of column ordering since the variables are bound in by name. I say partially because you still have to make sure that the columns & values portion of the insert are mapped correctly; but the dictionary itself can be in any order (since dicts are sort of unordered anyway)
There is probably a more pythonic way to achieve all this, but pulling the column names into a list and working off it ensures we have a static ordering to build the columns & values clauses.
data_dict = {'col1': 'value 1', 'col2': 'value 2', 'col3': 'value 3'}
columns = data_dict.keys()
cols_comma_separated = ', '.join(columns)
binds_comma_separated = ', '.join(['%(' + item + ')s' for item in columns])
sql = f'INSERT INTO yourtable ({cols_comma_separated}) VALUES ({binds_comma_separated})'
cur.execute(sql, data_dict)
Now whether or not it is a good idea to dynamically build your columns & values clause like this is a topic for a SQL injection thread.
This works for me
cursor.execute("INSERT INTO table (col) VALUES ( %(col_value) )",
{'col_value': 123})
if you have list in which there are number of dictionaries
for example: lst=[d1,d2,d3,d4]
then below one will worked for me:
for i in lst:
placeholders = ', '.join(['%s'] * len(i))
columns = ', '.join(i.keys())
sql = "INSERT INTO %s ( %s ) VALUES ( %s )" % (table, columns, placeholders)
cursor.execute(sql,list(i.values()))
conn.commit()
Note:Dont ever forget to commit otherwise you wont be able to see columns and values inserted in table
When constructing queries dynamically it’s important to ensure that both identifiers and values are correctly quoted. Otherwise you risk
- SQL injection if untrusted data is processed
- Errors if the column names require quoting (for example embedded spaces)
- Data corruption or errors if values are incorrectly quoted (for example
2021-07-11
unquoted may be evaluated as2003
)
Quoting values is best delegated to the DB-API connector. However connector packages don’t always provide a way to quote identifiers, so you may need to do this manually. MySQL uses backticks (`) to quote identifiers.
This code quotes identifiers and values. It works for MySQLdb, mysql.connector and pymysql and works for Python 3.5+.
data = {'col1': val1, 'col2': val2, ...}
# Compose a string of quoted column names
cols = ','.join([f'`{k}`' for k in data.keys()])
# Compose a string of placeholders for values
vals = ','.join(['%s'] * len(data))
# Create the SQL statement
stmt = f'INSERT INTO `tbl` ({cols}) VALUES ({vals})'
# Execute the statement, delegating the quoting of values to the connector
cur.execute(stmt, tuple(data.values()))
Let’s say our data is:
data = {
"name" : "fani",
"surname": "dogru",
"number" : 271990
}
This is my shorter version:
tablo = "table_name"
cols = ','.join([f" {k}" for k in data.keys()])
vals = ','.join([f"'{k}'" for k in data.values()])
stmt = f'INSERT INTO {tablo} ({cols}) VALUES ({vals})'